CVE-2026-85180

A flaw was found in Ollama. This vulnerability, known as Server-Side Request Forgery SSRF, occurs because Ollama fails to validate redirect destinations when pulling tensor-layer models. An unauthenticated attacker can exploit this by controlling a registry and serving a malicious tensor-layer...

Are AIs Still Struggling with CAPTCHAs?

Anthropic's recent security-incident document contains a bit about how CAPTCHAs are still frustrating Claude. In the transcript, the Claude model that is so powerful that Anthropic is gatekeeping access to it appeared to slam its virtual head against the wall solving a simple image identification...

CVE-2026-93764: Cleartext Storage of Sensitive Information

Mongoid may omit encryption rules for fields declared on embedded models when generating the client-side field-level encryption schema. Applications that enable this feature can therefore store values intended to be encrypted in readable form, with no error or warning. A party with routine read...

PT-2026-95563

Name of the Vulnerable Software and Affected Versions Mongoid affected versions not specified Description Mongoid may omit encryption rules for fields declared on embedded models during the generation of the client-side field-level encryption schema. This allows applications using this feature to...

EUVD-2026-81049

LightGBM through 4.7.0 fails to validate child and split array values when parsing text models, allowing attackers to write out-of-bounds memory during SHAP prediction. Attackers can craft malicious model files with invalid node references that trigger out-of-bounds writes at attacker-chosen...

CVE-2026-92221: Improper Neutralization of Special Elements used in an SQL Command

A vulnerability was determined in gedelumbung HospitalManagement up to c2d45543789a3887067d3915f69d44cfc2cf76a8. Affected by this vulnerability is the function generateindexpasien of the file application/models/appglobaladminmodel.php. Executing a manipulation of the argument cari can lead to sql...

CVE-2026-92786: Out-of-bounds Write

LightGBM through 4.7.0 fails to validate child and split array values when parsing text models, allowing attackers to write out-of-bounds memory during SHAP prediction. Attackers can craft malicious model files with invalid node references that trigger out-of-bounds writes at attacker-chosen...

PT-2026-94063

Name of the Vulnerable Software and Affected Versions LightGBM versions prior to 4.7.1 Description Failure to validate child and split array values when parsing text models allows for out-of-bounds memory writes during SHAP SHapley Additive exPlanations, a method used to explain the output of...

CVE-2026-91987 atomic-agents-stack before 1.1.0 Cost Guardrail Bypass via Unknown Model

atomic-agents-stack before 1.1.0 contains a cost-guardrail bypass in the estimatebatchcost function that returns zero cost for unknown models not in the pricing table. Attackers can configure deployments with unknown model identifiers to bypass daily cost caps and exceed budget limits in parallel...

CVE-2026-90842: Cleartext Storage in a File or on Disk

A weakness has been identified in PHPGurukul Blood Donor Management System 1.0. Affected by this issue is some unknown functionality of the file application/models/admin/LoginModel.php. This manipulation of the argument password/email/currentpassword/dbcurrentpwd/newpassword causes cleartext...

Vulnerability Localization Benchmark: Measuring Agentic Security Analysis at Repository Scale

Language-model agents increasingly operate over complete software repositories, yet cybersecurity evaluations primarily measure whether they can detect, reproduce, or repair vulnerabilities rather than whether they can locate the relevant code. We study vulnerability localization: given a weaknes...

Understanding the Usability of Cryptographic Verification Tools

Cryptographic protocol verification tools are widely used to analyze the security of complex protocols, yet how users interact with these tools remains comparatively understudied. We present an exploratory human-centered study of experienced users of Tamarin, ProVerif, and related protocol...

Permutation-Based Stegomalware in Large Language Models: Threats and Countermeasures

The difficulty of training large language models LLMs, together with their ubiquity, raises the threat of stegomalware, where malicious payloads are embedded into model weights. Recent work has demonstrated the use of permutation symmetry in model weights to mitigate these threats, but failed to...

Sociotechnical Aspects of Tor Relay Rejection

In 2019, the Tor Project enforced an end-of-life EoL policy for Tor versions, leading to the rejection of outdated relays, amounting to a notable fraction of consensus weight. While this policy aids network maintenance, reduces backporting efforts, and shortens vulnerability exposure, its...

Exploring Automated Vulnerability Identification in JavaScript Code Using Large Language Models

JavaScript powers approximately 98.8% of all websites, making vulnerabilities in its code a significant security risk, yet existing detection approaches such as Static Application Security Testing SAST tools often fail to identify many real-world vulnerabilities when applied to isolated code...

My Talk at DEF CON

Last month, I gave a talk at DEF CON on AI hacking: what happens when AIs become hackers. It's a combination of the potentialities I raised in my 2022 book A Hacker 's Mind and the lessons we're learning from current AI models engaging in hacking behavior. I'm really proud of the talk, and the fa...

The Models That Found 10,000 Zero-Days Broke Into Three Companies Using Weak Passwords

The question of whether a Frontier AI model could find vulnerabilities that no human researcher had found was settled in April. Claude Mythos Preview identified thousands of previously unknown flaws across every major operating system and browser, including a 27-year-old denial-of-service conditi...

CVE-2026-86753

snipe-it versions before 8.7.0 fail to validate the requestable flag for asset models in the POST /account/request/assetmodel/modelId endpoint. Authenticated users can bypass administrative restrictions and create checkout requests for non-requestable asset models by submitting requests directly ...

U.S. Agencies Accuse China AI Firms of Distilling Claude, GPT, Gemini, and Grok

U.S. cybersecurity and intelligence agencies have accused China-based artificial intelligence AI companies of conducting "systematic extraction" of proprietary functionalities and capabilities of American frontier models through distillation attacks. The activity has been described as occurring a...

PT-2026-89022

Name of the Vulnerable Software and Affected Versions snipe-it versions prior to 8.7.0 Description Authenticated users can bypass administrative restrictions to create checkout requests for asset models that are not marked as requestable. This occurs because the application fails to validate the...